Wikipedia talk:Citing sources
This is the talk page for discussing improvements to the Citing sources page. |
|
Archives: Index, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56Auto-archiving period: 75 days ![]() |
|
![]() | To find archives of this talk page, see this list. For talk archives from the previous Manual of Style (footnotes) page see Help talk:Footnotes. |
![]() | This project page does not require a rating on Wikipedia's content assessment scale. It is of interest to the following WikiProjects: | |||||||||||||||||||||||
|
What if I cite something in parentheses?
[edit]I encountered a situation today where I needed to put an additional detail in parentheses that is covered by a separate source from the sources used for the rest of the sentence. It seems kind of strange to use all three sources for the sentence that was there before both at the end and before the parentheses, but the source I added does not cover what came before the parentheses.— Vchimpanzee • talk • contributions • 19:44, 1 July 2024 (UTC)
- Would an explanatory foot note work? It will hide the content you are putting in parentheses until the reader clicks on the link, but it certainly makes the connection between the content and the reference clearer. If you want to keep the extra content always visible, you could also break up the sentence. Donald Albury 20:16, 1 July 2024 (UTC)
- Won't help if I don't know when I've done this previously. I found that one of the three sources for the entire sentence didn't verify anything, and got a 404 error for another source. So I concluded the third source would verify everything (it requires a subscription) and put it before what was in parentheses, and reworded so the information would match.— Vchimpanzee • talk • contributions • 22:24, 1 July 2024 (UTC)
Semi-protected edit request on 2 July 2024
[edit]![]() | This edit request has been answered. Set the |answered= or |ans= parameter to no to reactivate your request. |
Note: Not sure if my first attempt went through (please excuse if this is redundant).
Regarding "Slavery in colonial Spanish America" article:
4. Seijas, Tatiana.Asian Slaves in Colonial Mexico: From Chinos to Indians. New York: Cambridge University Press 2014.[page needed]
Add: space after author's name and "pp. 73-98" after the year of publication. Mearnest1 (talk) 19:01, 2 July 2024 (UTC)
- Not done. {{edit semi-protected}} is placed on a talk page to request an edit of the corresponding article or project. The citation mentioned in the request is not present on Wikipedia:Citing sources, rather, it is in Slavery in colonial Spanish America. That article does not appear to be protected. I'd do it myself but I don't have that book. Since Mearnest1 has done the research to find the page number, it's Mearnest1 who should make the edit. Jc3s5h (talk) 19:31, 2 July 2024 (UTC)
- TWL has the book 🤫 Folly Mox (talk) 20:42, 2 July 2024 (UTC)
Repeating publisher and location information for different articles from same website?
[edit]This is a relatively niche question. Let's say an article, such as AHS Krab, cites ten or more separate articles from the same news website. The citations cannot be combined using a single reference name, because each one links to a different URL. Must the publisher and location information be repeated for every single citation, or is it sufficient to include it in the first reference to that website? Huntthetroll (talk) 23:45, 4 July 2024 (UTC)
- I don't know if it will help in that case, but I did something like that in citing several sub-pages of a web site in Molasses Reef Wreck. Donald Albury 00:01, 5 July 2024 (UTC)
- (Sneaking in here so this reply makes sense) A similar effect can be achieved using {{harvc}}. Folly Mox (talk) 13:34, 5 July 2024 (UTC)
- @Huntthetroll, the editors at the article are welcome to set up whatever system they think is sensible. See WP:CITEVAR. WhatamIdoing (talk) 00:33, 5 July 2024 (UTC)
- I would make each cite stand totally by itself.
- In many articles references are removed or replaced with better ones. If that replace reference happened to be the one that contained the complete set of source details then they are lost for all the subsequent references from the same source. Of course, they are still in the article's history, so they could be recovered but at extra cost of editor effort - which often doesn't happen. Or sometimes the order of cites is changed, making a middle cite fuller then both preceding and following cites from the same source - weird looking!
- On the flip side, the cost of putting full the details in every reference from that source is just a copy/paste operation, so it is quite minimal effort. In fact, I often build up one in full by hand, then copy it many times and then alter the specific details - much quicker than typing it all by hand. Stepho talk 00:41, 5 July 2024 (UTC)
- I agree, making each citation self-contained seems best, as it is probably easiest to follow for the reader and robust in view of future changes. Gawaon (talk) 08:05, 5 July 2024 (UTC)
- Gawaon, I came to the same conclusion as you and |Stepho did when I was editing AHS Krab last night. Since I intend to continue adding content and citations to the article, I prioritize reader convenience and robustness in the face of a changing set of references. I will also investigate Folly Mox's suggestion about using {{harvc}}. Huntthetroll (talk) 17:23, 5 July 2024 (UTC)
- I agree, making each citation self-contained seems best, as it is probably easiest to follow for the reader and robust in view of future changes. Gawaon (talk) 08:05, 5 July 2024 (UTC)
- Huntthetroll, this reply will assume you're referring to the citations to the websites Defence24 and Altair, repeatedly cited with the respective parameters
|website= Defence24
|publisher= Defence24
|location= Warsaw, Poland
and|website= Altair
|publisher= Altair Agencja Lotnicza
|location= Warsaw, Poland
. I'd argue that the publisher and location of these websites are unnecessary in every case, including the first references to these sources.It's almost never helpful to include both|website=
and|publisher=
where the values for those parameters match to a large degree, as they do in these cases. It's also rare to include the|location=
of a website, unless it's the website of what used to be a physical news-paper. Folly Mox (talk) 13:29, 5 July 2024 (UTC)- I initially thought the same, but I eventually changed my mind.
- I tried to imagine the perspective of a reader that is completely unfamiliar with the article's topic, or with any of the sources cited. I would not expect the reader to know or assume that, for instance, the Polish-language website defence24.pl and the English-language website defence24.com are published by the same Polish limited-liability company, Defence24 Sp. z o.o. In fact, I was going to leave out the publisher for the similarly named site nowiny24.pl, because I assumed that the same company would be responsible, but decided to double-check the site's "O nas" ("about us") page, just to be sure. Suprisingly, nowiny24.pl is published by a completely different company, which should mean that it can be used to cross-check information from Defence24. I would not have known this, nor would any reader, if I had not looked up the publisher.
- In the case of a web citation, I treat
|location=
as the location of the publisher's headquarters. I find that this provides important information about the publisher's "institutional perspective", for lack of a better phrase, by showing the publisher's proximity to centers of political and economic power.
- Huntthetroll (talk) 17:52, 5 July 2024 (UTC)
- Sure! Whatever works best for your own editing flow and job satisfaction. I said above
I'd argue
, which appears to have been incorrect. Happy editing, Folly Mox (talk) 16:22, 6 July 2024 (UTC)
- Sure! Whatever works best for your own editing flow and job satisfaction. I said above
- I initially thought the same, but I eventually changed my mind.
Note, not ref
[edit]I messed up on The Iron Lady (film) and I can't find any explanation of what I should have done.— Vchimpanzee • talk • contributions • 16:58, 21 July 2024 (UTC)
- You're looking for {{efn}}, <note> doesn't do anything. See my edit[1]. -- LCU ActivelyDisinterested «@» °∆t° 17:40, 21 July 2024 (UTC)
- Thanks. I looked at an article that used notes and it didn't make sense what was done.— Vchimpanzee • talk • contributions • 17:55, 21 July 2024 (UTC)
- No worries, learning is part of the experience. -- LCU ActivelyDisinterested «@» °∆t° 18:00, 21 July 2024 (UTC)
- Thanks. I looked at an article that used notes and it didn't make sense what was done.— Vchimpanzee • talk • contributions • 17:55, 21 July 2024 (UTC)
WP:CITESHORT question
[edit]WP:CITESHORT says, (Note that templates should not be added without consensus to an article that already uses a consistent referencing style.)
Is this saying that I should not add citation templates to an article that does not already use citation templates? Schierbecker (talk) 17:42, 12 August 2024 (UTC)
- When in doubt, leave a note on the talk page. In practice, I don't remember ever being challenged when I have proposed changing citation style on an article. Also, many pages that have more than a couple of references already have a mixed style. Donald Albury 18:22, 12 August 2024 (UTC)
- It seems to be saying not to add templated citations to already cited pages, which is absurd. Further down the page it says
an article should not be switched between templated and non-templated citations without good reason and consensus
. If I'm rescuing a dead article with all the contributors long gone the first thing I'm doing is upgrading the refs with templates so that the short citations are followable to the long citations. I've never had a problem with this from other users. If we followed that rule, nearly every page created before ~2010 would still be using the legacy citation style (and we'd have a lot more dead links that the bots normally take care of when the refs are formatted as templates.) If I'm more comfortable adding citations by template then I should be able to do that. Schierbecker (talk) 20:02, 12 August 2024 (UTC)- In general you shouldn't change referencing styles, unless you are making extensive changes or rewriting the article. As ever "shouldn't" isn't the same as "mustn't", but if anyone objects you will need to find consensus to make the change before continuing. The issue is less one of absolutes, but rather about stopping editors from wasting their time arguing about what reference style to use. -- LCU ActivelyDisinterested «@» °∆t° 22:25, 12 August 2024 (UTC)
- On the other hand, many articles have a real mess of referencing, and a project of just cleaning up citations and creating a consistent style is justified. It is best to engage with regular editors on the talk page, if possible (see Talk:Joseph Conrad/Archive 2#Convert footnotes to Explanatory footnotes (efn)). Sometimes, nobody cares: Talk:Vaquita#Clean up needed - especially referencing. Donald Albury 23:26, 12 August 2024 (UTC)
- Yeah you shouldn't change an existing style, if there isn't one then imposing one style is considered helpful per 'Generally considered helpful' point 3 in WP:CITEVAR. -- LCU ActivelyDisinterested «@» °∆t° 10:49, 13 August 2024 (UTC)
- I think the fact that exceptions exist (e.g. for substantial rewrites, merges etc.) should be made more obvious. Overall, the admonishing against updating references is just spelled out too strongly. What about something along these lines:
- If untemplated references are preferred, take special care to maintain a consistent citation style throughout the article. Similarly, avoid changing templated citation styles without seeking consensus.
- Consistent citation styles are preferred. That being said, use whatever citation style you feel comfortable with. No one is required to know how to use your preferred citation style. If inconsistent citation styles bother you, fix it. Schierbecker (talk) 02:10, 13 August 2024 (UTC)
- The page already says that style should be consistent, and already included under "Generally considered helpful" is "making citations added by other editors match the existing style (if any). Do not revert someone else's contribution merely because the citation style doesn't match. If you know how to fix it, then fix it." Nikkimaria (talk) 02:17, 13 August 2024 (UTC)
- From the lead: While you should try to write citations correctly, what matters most is that you provide enough information to identify the source. Others will improve the formatting if needed.
- Maybe we should put it in bold.
- That said, I wonder if the community is ready to be done with the idea that "The use of citation templates is neither encouraged nor discouraged". WhatamIdoing (talk) 04:37, 21 August 2024 (UTC)
- On the other hand, many articles have a real mess of referencing, and a project of just cleaning up citations and creating a consistent style is justified. It is best to engage with regular editors on the talk page, if possible (see Talk:Joseph Conrad/Archive 2#Convert footnotes to Explanatory footnotes (efn)). Sometimes, nobody cares: Talk:Vaquita#Clean up needed - especially referencing. Donald Albury 23:26, 12 August 2024 (UTC)
- In general you shouldn't change referencing styles, unless you are making extensive changes or rewriting the article. As ever "shouldn't" isn't the same as "mustn't", but if anyone objects you will need to find consensus to make the change before continuing. The issue is less one of absolutes, but rather about stopping editors from wasting their time arguing about what reference style to use. -- LCU ActivelyDisinterested «@» °∆t° 22:25, 12 August 2024 (UTC)
- It seems to be saying not to add templated citations to already cited pages, which is absurd. Further down the page it says
Source is not the newspaper, but it looks like source has a page number
[edit]This is the diff. Notice that when you look under references, the page number from the newspaper makes it look like it is a page number in the original source.
Also, I should point out that I can't create a clip, so access to the source is currently limited to Wikipedia library users.— Vchimpanzee • talk • contributions • 21:40, 20 August 2024 (UTC)
- Not really clear what you are asking, if you are asking anything. Presumably, you have read the source linked by the url. You have actually consulted this source, right? The page number is in the bottom left margin. If you have not, then you shouldn't be using that source in an en.wiki article.
- Do not use https://www-newspapers-com.wikipedialibrary.idm.oclc.org/ urls. To do so, does a disservice to readers who aren't editors because they will never get beyond the Wikipedia Library banner page. Use the correct newspapers.com url. There is some discussion about clipping at WP:Newspapers.com.
- —Trappist the monk (talk) 22:06, 20 August 2024 (UTC)
- The page number is the page number in the Concord Monitor. But if you look at the ref, it looks like it is a page number in TV Media, which provided the article to the Concord Monitor.
- I'm not sure how to convert the newspapers.com URL because of the problem that I linked to.— Vchimpanzee • talk • contributions • 22:09, 20 August 2024 (UTC)
- I followed the bot directions and did get the URLs converted. So is the page number all right the way it is?— Vchimpanzee • talk • contributions • 22:25, 20 August 2024 (UTC)
- Use the page number as it is written in the source.
- —Trappist the monk (talk) 22:30, 20 August 2024 (UTC)
- (edit conflict)
- Look in the lower left margin. It gives both Concord Monitor and Sunday TV Magazine names along with the page number and issue date. You could write a
{{cite magazine}}
with|magazine=Sunday TV Magazine
and|via=[[Newspapers.com]] / ''[[Concord Monitor]]''
or some such. - Apparently, others have solved the clipping issue. That is why I linked to WP:Newspapers.com. If you have questions about clipping, you should ask at the WP:Newspapers.com talk page. If the current url cannot be translated, remove it.
- —Trappist the monk (talk) 22:30, 20 August 2024 (UTC)
- The page number is not the page number of Sunday TV Magazine, but the TV Magazine of the Concord Monitor.
- For the clipping, I'm not doing whatever they did. If someone wants to create a clipping using my ref, they can, because I did fix the link. I'm waiting until I can create a clipping myself without doing something exceedingly complicated.— Vchimpanzee • talk • contributions • 17:50, 21 August 2024 (UTC)
- I followed the bot directions and did get the URLs converted. So is the page number all right the way it is?— Vchimpanzee • talk • contributions • 22:25, 20 August 2024 (UTC)
- Is this just a WP:SAYWHEREYOUGOTIT problem? WhatamIdoing (talk) 04:39, 21 August 2024 (UTC)
- I didn't mean to get into the issue of how to link to newspapers.com. The problem is not that, but how to make it clear what the page number refers to. It looks like it is the page number of "TV Media".— Vchimpanzee • talk • contributions • 17:43, 21 August 2024 (UTC)
Order of the example in WP:SAYWHERE
[edit]Maybe this is really minor. Currently, WP:SAYWHERE has the order CITED by CITING. Should not it take the order CITING citing CITED? That is, in Smith (2009), p. 99, cited in Jones (2010), p. 29.
instead Jones (2010), p. 29, citing Smith (2009), p. 99.
That would place the actual location of the material first. Ifly6 (talk) 17:54, 27 August 2024 (UTC)
- I think I'd put the source that you personally read yourself first, but either is probably okay. WhatamIdoing (talk) 23:35, 30 August 2024 (UTC)
- I'm used to the "cited in" order given in the example. And like WhatamIdoing says, either should be fine, so I don't see a reason for a change. "Actual location" doesn't really apply, since the quoted text should be present in both locations. And the "CITED first" order had the advantage of crediting the original/actual author first. Gawaon (talk) 05:33, 31 August 2024 (UTC)
- I'm too lazy to check, but I think that both options used to be present. Personally I prefer the one that is given now, but both are acceptable. Zerotalk 07:31, 31 August 2024 (UTC)
sharing Zotero libraries with other wiki-editors
[edit]I have several Zotero and EndNote libraries with detailed bibliography and full-text pdfs of Open Access publications. The topics cover lithium-ion batteries, sodium-ion batteries, flow batteries, international order, nuclear warfare, nuclear submarines, persistent organic pollutants etc. The data came from Scopus, Web of Science, The Lens, CORE (research service) and other databases. I would like to share these libraries with interested wiki-editors, and I wonder if Wikipedia has a mechanism for such sharing. Walter Tau (talk) 15:37, 30 August 2024 (UTC)
- @Mvolz, are you around? WhatamIdoing (talk) 23:36, 30 August 2024 (UTC)
- The only place we centrally share structured citation is as wikidata items and I don't think there's an easy way to import from those libraries to wikidata. Mvolz (talk) 08:51, 31 August 2024 (UTC)
- Thank you for reminding me about WikiData. I do not think I need to export anything from my libraries.
- I can just post a Zotero file (or its archived version) on to WikiData.
- Let me give it a try and see how it works.
- Have a good day. Walter Tau (talk) 10:09, 31 August 2024 (UTC)
- Wikidata is not meant for arbitrary files. Gawaon (talk) 11:10, 31 August 2024 (UTC)
- Dear Gawaon: thank you for your comment. Is there is a decription of what is suitable for wikidata and what is not? Also, am I the very first wikipedian, who wants to share a searcheable database (with or without full texts) with others? I would think, that many wiki-articles (or topics) would have such databases by now (especially, if they are created using no-restrictions sources like The Lense. If no such option exists today, how can I post it for a discussion? OR would you be willing to do it, since you may know better how such thins work here? Walter Tau (talk) 15:53, 31 August 2024 (UTC)
- Wikidata has citation/bibliographic information for a large number of scientific papers. I think that Daniel Mietchen has done some of the work on that.
- (For myself, I keep wishing someone will do some mw:Citoid/Creating Zotero translators for the BBC's website. It always surprises me that the visual editor doesn't recognize it as a news site, and can't pull most of the information.) WhatamIdoing (talk) 17:03, 31 August 2024 (UTC)
- Dear Gawaon: thank you for your comment. Is there is a decription of what is suitable for wikidata and what is not? Also, am I the very first wikipedian, who wants to share a searcheable database (with or without full texts) with others? I would think, that many wiki-articles (or topics) would have such databases by now (especially, if they are created using no-restrictions sources like The Lense. If no such option exists today, how can I post it for a discussion? OR would you be willing to do it, since you may know better how such thins work here? Walter Tau (talk) 15:53, 31 August 2024 (UTC)
- Wikidata is not meant for arbitrary files. Gawaon (talk) 11:10, 31 August 2024 (UTC)
Videos
[edit]Can someone please make a series of YouTube videos going through and verbally explaining, with examples everything on the Wikipedia help pages?
Thank you very much for your time and consideration.
I look forward to hearing from you soon. Cole Massi1 (talk) 18:50, 4 September 2024 (UTC)
- @Cole Massi1: Have you searched YouTube? There are a whole bunch of solid introduction videos there. Rjjiii (talk) 03:30, 5 September 2024 (UTC)
citation generator?
[edit]What's the easy way for a new user to generate citations from archive.org or google.books for the 25 books listed on Sydney_Moseley#Works? Enri999 (talk) 05:32, 9 September 2024 (UTC)
- @Enri999, there isn't an automatic way to do that correctly. You will have to look up each one of them manually. The convention is to use the first edition for such lists.
Auto-citing a source using the visual editor – look for the "Cite" button in the toolbar. - Once you have the URLs, here's an example of what you can expect as an autogenerated ref in the visual editor (though you'll have to switch to wikitext to remove the ref tags), using the first book:
- Moseley, Sydney A. (Sydney Alexander) (1916). The truth about the Dardanelles. Robarts - University of Toronto. London : Cassell.
- Moseley, Sydney Alexander (1916). The Truth about the Dardanelles. Cassell, Limited.
- The first is archive.org and the second is books.google.com. (You can also generate citations from an ISBN, but these are all too old for that.) After they're generated, you can edit it to change anything that you think it got wrong. If you've not tried the visual editor, then this link will probably work for you: https://en.wikipedia.org/wiki/Sydney_Moseley?veaction=edit It works like Google Docs. WhatamIdoing (talk) 05:49, 9 September 2024 (UTC)
- Thank you, I didn't know about that button, that will help. Cheers. Enri999 (talk) 05:54, 9 September 2024 (UTC)
- @Enri999; WhatamIdoing: And like so, so, many automatically created citations, those are both flawed.
|first=Sydney A. (Sydney Alexander)
– don't do that; write the name as it appears in the source:|first=Sydney A.
|publisher=London : Cassell
– don't do that; write the name as it appears in the source:|location=London
|publisher=Cassell and Company
|others=Robarts - University of Toronto
– don't do that; the name of the institute that contributed the source to Internet Archive is irrelevant and may mislead our readers|first=Sydney Alexander
– don't do that; write the name as it appears in the source:|first=Sydney A.
|publisher=Cassell, Limited
– don't do that; omit corporate designations unless required for disambiguation:|publisher=Cassell and Company
- You cannot trust visual editor/citoid to auto-magically create correct citations; they are dependent on the quality of the metadata that can be scraped from whatever online source. Sure, use ve/citoid to fetch some of the source's metadata but you must check and correct each and every citation that the tool creates. Be responsible and don't create a mess that other editors will have to clean up.
- —Trappist the monk (talk) 07:23, 9 September 2024 (UTC)
- The first one also has the title in librarian-preferred Sentence case, even though most citations use Title case. No autogenerating system is going to turn bad metadata into perfect citations, and I think this is the level of imperfection that you can realistically expect. WhatamIdoing (talk) 07:37, 9 September 2024 (UTC)
- That's probably true without a lot of postprocessing, and the initial algorithmic citations are still pretty bad. I'm not sure what all can be done to make it clear that any citation generation script is a first-pass tool that will almost always produce output requiring manual adjustment. I'm hoping Edit Check might help eventually, as would surfacing CS1 maint messages within the VE interface, as would additing a translation layer on top of Citoid, adding lots and lots of special cases, etc. I personally find that automatically generated citations typically require so much tweaking that it's generally not a timesave even to begin with them unless there are more than eight or nine authors. Folly Mox (talk) 22:00, 9 September 2024 (UTC)
- I guess there is a reason why I just plod along building citations manually. Donald Albury 22:47, 9 September 2024 (UTC)
- @Folly Mox: What is "Edit Check"? Rjjiii (talk) 03:25, 10 September 2024 (UTC)
- See mw:Edit check. The first "check" is encouraging new editors to add citations, if they are adding a new paragraph. (Of course there are other times when adding a citation would be appropriate, but it's an easy-ish thing for the software to detect, and it's almost never a bad idea to add a citation if you write, e.g., a whole new paragraph.) WhatamIdoing (talk) 04:21, 10 September 2024 (UTC)
- I have high hopes for future extended functionality that can provide realtime feedback about mistakes and problematic edits. Something a bit more nuanced and informative than edit filters. Rjjiii, it's not currently implemented on English Wikipedia at all, and the stretch goals are yet but dreams. Folly Mox (talk) 04:53, 10 September 2024 (UTC)
- I have my doubts about whether Clippy would be welcome, but there are a few things that could be handy (e.g., pre-warning about URLs that are on the Wikipedia:Spam blacklist). WhatamIdoing (talk) 05:15, 10 September 2024 (UTC)
- Right, nobody wants Wikipe-tan intruding into their editing interface all Comic Sans "Looks like you're populating an infobox! Do you want to navigate away to a tangentially related Help: page instead?" I'm conceptualising the feature I've done exactly zero work on as more like a fully automated Twinkle, dropping boilerplate modals at rookie mistakes that established contributors tend over time towards becoming increasingly frustrated and bitey about.Our documentation is... not really presented in a way that minimises common errors for newer editors. Presenting applicable guidance on an as-needed basis feels like it should be mostly positive. Folly Mox (talk) 06:31, 10 September 2024 (UTC)
- One of the lessons from the original Clippy was that newbies appreciated its assistance only for a very brief period of time. When you had just bought your first-ever computer, at a cost of almost a month's income, any friendly-looking help was appreciated, especially if you'd never used a computer before, and might not even know how to type.
- But after the first jitters wore off, most people learned how to use their new computers quickly, and they equally quickly wanted to get rid of anything that treated them like a newbie.
- With that in mind, it's possible that we should design for universal use (e.g., autogenerated refs, because even though they're imperfect, they are very popular with editors of all experience levels), or for bots that autofix the rookie mistakes (e.g., we don't have to revert newbies dropping Facebook links into articles, because XLinkBot does it faster than humans can). WhatamIdoing (talk) 07:49, 10 September 2024 (UTC)
- I agree (I think you're advocating this; please correct me if not) that improving Citoid's output (in addition to that of scripts like reFill and Citation bot) is probably a more fruitful avenue for raising the quality of citations added across experience levels. I think this starts with some sort of community configured functionality that hooks into Citoid or VE somewhere, which there was a subthread about at the recent VPR thread on Edit Check.As stated above,
No autogenerating system is going to turn bad metadata into perfect citations
, so for a "full solution", one necessity seems to be some sort of acculturation into the practice of double checking code output rather than blindly trusting it.As to the other issue we appear to have differing perspectives on – the hypothetical future usefulness of / annoyance with potential Edit Check cases – I'm not sure if we actually disagree or if we're not understanding each other's prior assumptions.I feel like I would derive value as an editor from a script that warned me if I e.g. left a date or copypasted superscript numeral in an author name field, similar to the thing that prompts me if my edit summary is blank (which doesn't function in Minerva). And I would derive value as a patroller from a process that prompted people e.g. not to change a shortdesc manually set tonone
into a vacuous / pleonastic synonym of the article title because they think having no short description is erroneous and are unaware of WP:SDNONE, or altering variant English spellings to their preferred lect's because they're unaware of valid alternatives and MOS:ENGVAR, or altering era styles from BCE/CE to BC/AD or vice versa because they have feelings about it and are unaware of MOS:ERA, or altering ordinals from English words to numeric representations because they're unaware of MOS:ORDINAL, etc.All of these are pretty common, typically reverted, and not bot-addressible because they'd all fall under WP:CONTEXTBOT, and could be the valid result of a talkpage discussion, conformance with a "Use Regional English" template, or something similar, although the likelihood is low.Having thought on it a bit instead of getting ready for work, I suppose the initial Edit Check message about adding a reference to new uncited paragraphs might be encountered frequently enough to generate annoyance in users (and maybe the idea of warning for impending 3rr violation, although a warning for 1rr on affected articles would probably be more valuable), but most of the things I'm envisioning should probably display once, create a moment of education, and then not be triggered again for the same user unless they are disruptively editing against consensus practice, in which case we should want to annoy them, which might help them stop without technical restriction. Folly Mox (talk) 11:25, 10 September 2024 (UTC)
- I agree (I think you're advocating this; please correct me if not) that improving Citoid's output (in addition to that of scripts like reFill and Citation bot) is probably a more fruitful avenue for raising the quality of citations added across experience levels. I think this starts with some sort of community configured functionality that hooks into Citoid or VE somewhere, which there was a subthread about at the recent VPR thread on Edit Check.As stated above,
- Right, nobody wants Wikipe-tan intruding into their editing interface all Comic Sans "Looks like you're populating an infobox! Do you want to navigate away to a tangentially related Help: page instead?" I'm conceptualising the feature I've done exactly zero work on as more like a fully automated Twinkle, dropping boilerplate modals at rookie mistakes that established contributors tend over time towards becoming increasingly frustrated and bitey about.Our documentation is... not really presented in a way that minimises common errors for newer editors. Presenting applicable guidance on an as-needed basis feels like it should be mostly positive. Folly Mox (talk) 06:31, 10 September 2024 (UTC)
- I have my doubts about whether Clippy would be welcome, but there are a few things that could be handy (e.g., pre-warning about URLs that are on the Wikipedia:Spam blacklist). WhatamIdoing (talk) 05:15, 10 September 2024 (UTC)
- I have high hopes for future extended functionality that can provide realtime feedback about mistakes and problematic edits. Something a bit more nuanced and informative than edit filters. Rjjiii, it's not currently implemented on English Wikipedia at all, and the stretch goals are yet but dreams. Folly Mox (talk) 04:53, 10 September 2024 (UTC)
- See mw:Edit check. The first "check" is encouraging new editors to add citations, if they are adding a new paragraph. (Of course there are other times when adding a citation would be appropriate, but it's an easy-ish thing for the software to detect, and it's almost never a bad idea to add a citation if you write, e.g., a whole new paragraph.) WhatamIdoing (talk) 04:21, 10 September 2024 (UTC)
- That's probably true without a lot of postprocessing, and the initial algorithmic citations are still pretty bad. I'm not sure what all can be done to make it clear that any citation generation script is a first-pass tool that will almost always produce output requiring manual adjustment. I'm hoping Edit Check might help eventually, as would surfacing CS1 maint messages within the VE interface, as would additing a translation layer on top of Citoid, adding lots and lots of special cases, etc. I personally find that automatically generated citations typically require so much tweaking that it's generally not a timesave even to begin with them unless there are more than eight or nine authors. Folly Mox (talk) 22:00, 9 September 2024 (UTC)
- The first one also has the title in librarian-preferred Sentence case, even though most citations use Title case. No autogenerating system is going to turn bad metadata into perfect citations, and I think this is the level of imperfection that you can realistically expect. WhatamIdoing (talk) 07:37, 9 September 2024 (UTC)